Supported rhodium nanoparticles in catalysis: The role of stabilizers on catalytic activity and structural features

Article abstract of DOI:10.1016/S0022-328X(03)00529-1

Rhodium nanoparticles supported on γ-Al₂O₃, derived from arene-solvated Rh atoms stabilized by trioctylamine (TOA), are valuable catalysts in hydrogenation and silylformylation reactions. They are largely more active than the analogous commercial catalyst, as well as of a sample similarly prepared in the absence of TOA. HRTEM measurements, IR studies on adsorbed CO species and extended X-ray absorption fine structure analyses evidenced the role of TOA in controlling the particle growth in the preparative process and in stabilizing the resulting Rh particles against erosion by CO and oxidation.

Full text of DOI:10.1016/S0022-328X(03)00529-1

Journal of Organometallic Chemistry 681 (2003) 37Á50
/
www.elsevier.com/locate/jorganchem
Supported rhodium nanoparticles in catalysis: the role of stabilizers
on catalytic activity and structural features
a,
Giovanni Vitulli *, Claudio Evangelisti , Paolo Pertici , Anna Maria Caporusso ,
a
a
a
a
Nicoletta Panziera , Piero Salvadori , Maria Giulia Faga , Chiara Manfredotti ,
a
b
b
b
Gianmario Martra , Salvatore Coluccia , Antonella Balerna , Stefano Colonna ,
b
c
d
c,e
Settimio Mobilio
a
Istituto per la Chimica dei Composti OrganoMetallici, ICCOM-CNR, Sezione di Pisa, c/o Dipartimento di Chimica e Chimica Industriale
dell’Universit a` di Pisa, via Risorgimento 35, 56126 Pisa, Italy
Dipartimento di Chimica I.F.M., Universit a` di Torino, via P. Giuria 7, 10125 Torino, Italy
b
c
INFN-Laboratori Nazionali di Frascati, Via E. Fermi 40, 00044 Frascati, Roma, Italy
d
CNR-Istituto di Struttura della Materia, Via Fosso del Cavaliere 100, 00133 Roma, Italy
e
Dipartimento di Fisica, Universita’ di Roma Tre, Via della Vasca Navale 84, 00146 Roma, Italy
Received 10 March 2003; received in revised form 8 May 2003; accepted 8 May 2003
Abstract
Rhodium nanoparticles supported on g-Al O , derived from arene-solvated Rh atoms stabilized by trioctylamine (TOA), are
2
3
valuable catalysts in hydrogenation and silylformylation reactions. They are largely more active than the analogous commercial
catalyst, as well as of a sample similarly prepared in the absence of TOA. HRTEM measurements, IR studies on adsorbed CO
species and extended X-ray absorption fine structure analyses evidenced the role of TOA in controlling the particle growth in the
preparative process and in stabilizing the resulting Rh particles against erosion by CO and oxidation.
#
2003 Elsevier Science B.V. All rights reserved.
Keywords: Supported rhodium nanoparticles; Trioctylamine; Metal vapour synthesis; Catalysis; Surface properties
1
. Introduction
hydrogenation reaction, for instance, strong interacting
ligands, like tetraalkylammonium halides, poly-N-vinyl-
2-pyrrolidone, 1,10-phenantroline, have been found to
restrict the catalytic activity [1j], while hydroxyalkyl
ammonium salts or aliphatic amines with long alkyl
Recently, there has been a great deal of attention on
the use of metal nanoparticles as valuable catalysts [1].
Their catalytic properties are strongly related to the
particle size and recent findings on their size-dependent
chemistry has added a new dimension to the subject [2].
However, stabilizers are usually used to avoid agglom-
eration to bulk metal and the kind of stabilizer can also
strongly affect their catalytic behaviour [3]. In the
groups (C or higher) have been reported to greatly
8
improve the efficiency of the catalytic precursor [1i,4].
Moreover, solvent-stabilized metal microclusters, ob-
tained by reaction of metal vapour with weakly co-
ordinating organic ligands (acetone, tetrahydrofuran,
toluene, etc.) (solvated metal atoms), are very active
homogeneous catalytic precursors in a wide range of
reactions, including hydrogenation, hydrosilylation, hy-
droformylation, silylformylation and oligo-polymeriza-
tion of olefins and acetylenes [1c,1l,5]. With the aim to
provide a further insight into the subject, we have
*
Corresponding author. Present address: Istituto per la Chimica dei
Composti Organo Metallici, ICCOM-CNR, Sezione di Pisa, via
Risorgimento 35, Pisa 56125, Italy. Tel.: ꢀ39-050-918224; fax: ꢀ39-
/
/
050-918260.
E-mail address: pervit@server1.dcci.unipi.it (G. Vitulli).
0022-328X/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0022-328X(03)00529-1

38
G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á50
/
studied the activity of catalysts prepared from arene-
solvated Rh atoms, containing trioctylamine (TOA) as
additional stabilizer. Such systems are excellent starting
materials for the preparation of new supported catalysts
largely more active and selective than analogous samples
commercially available or similarly prepared in the
absence of TOA. We report here evidence for their
peculiar catalytic properties and some data on their
structural features. For this, the catalysts have been
characterized by HRTEM analyses, IR studies on
adsorbed CO species and extended X-ray absorption
fine structure (EXAFS) investigations.
the micrographs at least 300 particles; the mean particle
diameter (d ) was calculated by using the formula d ꢂ
/
m
m
ad n /an where ni was the number of particles of
i
i
i
diameter d . In the case of metal particles of the Rh/g-
i
Al O (MVS) systems, the diameter d used for the
calculation in these two cases has been the medium value
of each dimensions interval in the histogram, thus with a
2
3
i
resulting uncertainty of 90.25 nm. The reason for this
/
different data treatment is that rhodium particles in
these two cases were very small and the usual mean
diameter calculation (where the d considered is usually
i
the highest limit of each dimensions interval) lead to an
uncertainty of 90.5 nm for the mean diameter obtained,
/
and such uncertainty limit was too high for the
considered metal particles.
2
. Experimental
The catalytic samples were prepared for the EXAFS
experiments by suspending the fine powders mixed with
boron-nitride in ethanol, drying the mixtures under
vacuum and pressing them into pellets. The sample
amount was calculated in order to achieve a reasonable
edge jump of about 0.7.
The EXAFS measurements at the Rh K-edge (23.220
keV) have been performed at the GILDA beamline [7]
of the European Synchrotron Radiation Facility
(ESRF, Grenoble). The radiation was monochromated
and horizontally focused by using a double crystal
Si(3 1 1) monochromator [8]; two Pt mirrors were used
to reject the higher harmonics and focus the beam in the
vertical direction [7].
2
.1. General
All operations involving the metal vapour synthesis
MVS) products were performed under a dry argon
(
atmosphere. The co-condensation of rhodium and
mesitylene was carried out in a static reactor previously
described [6]. The ‘‘mesitylene-solvated Rh atoms’’
solution was worked up under argon atmosphere with
the use of the standard Schlenk techniques. The amount
of rhodium in the above solutions was determined by
atomic absorption spectrometry in a electrochemically
heated graphite furnace with a PerkinÁElmer 4100ZL
/
instrument. Solvents were purified by conventional
methods, distilled and stored under argon. TOA,
cinnamaldehyde, methyl benzoate, 1-hexyne were com-
mercial products and were degassed and stored under
argon before use. Commercial g-Al O (Chimet product,
Two reference samples (a-Rh O and Rh foil) and
2
3
two MVS derived catalysts (Rh/g-Al O and Rh(TOA)/
2
3
g-Al O ) were studied, all in transmission mode, using
2
3
two Ar-filled ionization chambers to detect the incident
(I ) and transmitted (I ) X-ray beam. The measured
2
3
2
ꢁ1
type 49, surface area 110 m g , mean particle diameter
0
T
3
1 mm) was dried in a static oven before use. Commer-
absorption coefficient is defined as m(E)ꢂln(I /I ). All
spectra were recorded at 77 K in order to reduce the
thermal effects.
/
0 T
cial rhodium on g-Al O (5 wt.% of Rh, surface area 130
2
3
2
m g ) was an Engelhardt product.
ꢁ1
The infrared spectra of adsorbed CO studies were
obtained with a Bruker IFS-28 spectrometer with MCT
The GLC analyses were performed on a PerkinÁ
/
Elmer Auto System gas chromatograph, equipped with
a flame ionization detector (FID), using a SiO column
ꢁ
1
detector (128 scans, resolution 4 cm ). H and CO with
2
2
high-purity grade were used without further purifica-
tion. The samples, in the form of self-supporting pellet,
were put into an IR quartz cell with KBr windows and
outgassed at room temperature (r.t.) for 50 min (residual
(DB1, 30 mꢃ
/
0.52 mm, 5 mm) and helium as carrier gas.
1
13
H- and C-NMR spectra were measured on Varian
Gemini 200 spectrometer at 200 and 50.3 MHz,
respectively, using chloroform-d as solvent; chemical
shifts are relative to internal Si(CH ) .
ꢁ
5
pressure into the vacuum line B10 Torr). After that,
/
3
4
the samples were treated with H (100 Torr) for 12 h at
2
r.t., outgassed at r.t. for 1 h and then put into contact
with 100 Torr of CO, monitoring the spectra at
increasing contact time with CO up to 2 h.
2.2. Preparation of the rhodium catalysts
2.2.1. Synthesis of the solvated Rh atoms
In a typical experiment, rhodium vapour, generated
by resistive heating of a tungsten wire surface coated
Electron micrographs were obtained by a Jeol
000EX microscope. Before the introduction in the
2
instrument, the samples, in the form of powders, were
ultrasonically dispersed in isopropyl alcohol and a drop
of the suspension was deposited on a copper grid
covered with a lacey carbon film. Histograms of the
particle size distribution were obtained by counting onto
with electrodeposited rhodium (100Á150 mg), was co-
/
condensed at liquid nitrogen temperature with mesity-
lene (40 ml) in the glass reactor chamber of the MVS
apparatus in ca. 40 min. The reactor chamber was
warmed to the melting point of the solid matrix (ca. ꢁ
/

G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á
/50
39
4
0 8C) and the resulting brown solution was siphoned at
low temperature in a Schlenk tube and kept in a
refrigerator at ꢁ20 8C. The content of the metal
solution was 1.2 mg Rh ml
analysis of liquid samples removed by the tap. The
methyl cyclohexenoate was identified by comparison of
its GLC retention time with that of an authentic sample.
The conversion as function of the time for different
catalysts is reported in Fig. 1.
/
ꢁ
1
ꢁ1
(0.0117 mg atom ml ).
2
.2.2. Preparation of rhodium on g-Al O catalysts
2 3
The above brown Rh/mesitylene reaction solution
mesitylene-solvated metal atoms) (8.34 ml, 10 mg Rh)
was added to a suspension of g-Al O (10 g) in
(
2.3.2. Hydrogenation of cinnamaldehyde: general
procedure
2
3
mesitylene (20 ml). The mixture was stirred for 12 h at
room temperature. The colourless solution was removed
and the light-brown solid, containing 0.1 wt.% Rh, was
washed with pentane and dried under reduced pressure.
The Rh/g-Al O catalysts, containing 1 and 5 wt.% Rh,
A 50-ml round-bottomed flask was charged with the
rhodium catalyst (0.48 mg atom Rh), toluene (15 ml)
and cinnamaldehyde (6.1 ml, 48 mmol). The mixture
was stirred (ca. 1200 rpm) at room temperature for 12 h.
The solution was separated from the catalyst by
decantation and analysed by GLC. The products were
identified by comparison of their GLC retention times
with those of authentic samples. The results obtained
using different catalysts are reported in Table 1.
2
3
were prepared similarly using 1 g and 190 mg of g-
Al O , respectively.
2
3
2
g-Al O catalysts
.2.3. Preparation of rhodiumÁ
/trioctylamine (TOA) on
2
3
2
.2.3.1. Preparation of Rh(TOA)/mesitylene solution.
TOA (0.30 ml, 0.69 mmol) was added at the MVS Rh/
mesitylene solution (20 ml, 0.23 mg atom) and the
mixture was stirred for 1 h at room temperature. The
resulting brown Rh(TOA)/mesitylene solution was
stable at room temperature for a long time (ca. 3
months).
2
.3.3. Silylformylation of 1-hexyne: general procedure
The silylformylation experiments were carried out at
0 atm CO starting pressure and room temperature in a
5-ml stainless steel autoclave previously described (see
1
2
Section 2.3.1). In a typical experiment, the autoclave
was charged under argon atmosphere with the rhodium
catalyst (0.002 mg atom), toluene (2 ml), dimethylphe-
nylsilane (0.31 ml, 2 mmol) and 1-hexyne (0.23 ml, 2
mmol). The autoclave was closed, argon was eliminated
and carbon monoxide was introduced (10 atm). The
reaction mixture was stirred (1200 rpm) and monitored
by GLC analysis of liquid samples removed by the tap.
When silane was completely converted, the autoclave
was discharged, pentane (20 ml) was added and the
surnatant liquid was separated from the catalyst by
decantation. The volatile materials were removed under
vacuum and the raw product was recovered by bulb-to-
bulb distillation (1 mmHg). Pure (Z)-1-(dimethylphe-
nylsilyl)-2-formyl-1-hexene was obtained by column
chromatography (silica gel; eluent, pentane:ethylacetate,
2.2.3.2. Preparation of the Rh(TOA)/g-Al O catalysts.
2 3
The brown Rh(TOA)/mesitylene solution (8.34 ml, 10
mg Rh) was added to a suspension of g-Al O (10 g) in
2
3
mesitylene (20 ml). The mixture was stirred for 12 h at
room temperature. The colourless mesitylene was re-
moved and the light-brown solid, containing 0.1 wt.%
Rh, was washed with pentane and dried under reduced
pressure. The Rh/g-Al O catalysts, containing 1 and 5
2
3
wt.% Rh, were prepared similarly using 1 g and 190 mg
of g-Al O , respectively.
2
3
2
.3. Catalytic reactions
2.3.1. Hydrogenation of methylbenzoate: general
procedure
1
13
95/5) and characterized by H- and C-NMR [1j].
The hydrogenation experiments were performed at 20
atm starting hydrogen pressure using a 75-ml stainless
steel autoclave equipped with a Teflon removable
cylinder Jacker, a magnetic bar, a stainless steel tap
and a manometer (100 atm end scale). The reactions
were performed by heating the autoclave in an oil bath
1
a
H-NMR (200 MHz) d: 0.51 (6H, H , s), 0.91 (3H,
b
bc
c
d
H , t, J ꢂ
/
7.0 Hz), 1.22Á1.45 (4H, m, H , H ), 2.31
/
at 8091 8C. In a typical experiment, the autoclave was
/
e
ed
f
(
2H, H , t, J ꢂ
(3H, H , m), 7.48Á
ppm. C-NMR (50.3 MHz) d: ꢁ0.3, 13.7, 22.3, 30.4,
31.4, 128.3, 129.6, 133.7, 137.2, 149.2, 157.4, 193.7 ppm.
The results with different catalysts are reported in
Table 2.
/
7.3 Hz), 6.93 (1H, H , s), 7.35Á7.40
/
charged, under argon atmosphere, with the rhodium
catalyst (0.064 mg atom), THF (10 ml) and methyl-
benzoate (0.8 ml, 6.4 mmol). The autoclave was closed,
argon was eliminated and hydrogen was introduced (20
atm). The reaction mixture was stirred (1200 rpm) and
heated to 80 8C. The reaction was monitored by GLC
g
g
7.55 (2H, H , m), 9.78 (1H, H , s)
h
/
1
3
/

40
G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á50
/
Fig. 1. Hydrogenation of methyl benzoate to methyl cyclohexanoate with rhodium on g-alumina catalysts.
3
. Results and discussion
which gives, on melting, a brown solution stable at
40 8C containing mesitylene-stabilized Rh-hydride
microclusters [9]. The warming up to room temperature
ꢁ
/
3
.1. Preparation of supported Rh catalysts
in the presence of supports, such as g-Al O , results, by
2
3
The reaction of Rh vapour with mesitylene in large
excess, at liquid nitrogen temperature in a reactor
elsewhere described [6], affords a reddish-brown matrix
a further clustering process, in a gentle deposition of
well-dispersed Rh nanoparticles on the support [10]
(Scheme 1).
Table 1
a
2 3
Hydrogenation of cinnamaldehyde with rhodium on g-Al O catalysts
Catalyst
Conversion (%)
Product distribution (%)
b
1
2
3
Other
Rh/g-Al
Rh/g-Al
Rh(TOA)/g-Al
Rh(TOA)/g-Al
2
O
O
3
(1%)
(0.1%)
40
5
5
25
60
25
10
75
10
10
2
3
Á/
Á/
2
O
O
3
(1%)
(0.1%)
82
75
5
Á/
Á/
80
75
80
10
15
10
2
3
Rh/g-Al
2
O
3
(5%) commercial
10
Á/
a
Reaction conditions: Rh catalyst, 0.48 mg atom; cinnamaldehyde, 6.1 ml (48 mmol); toluene, 15 ml; atmospheric hydrogen pressure; 25 8C;
reaction time, 12 h.
b
Unidentified high boiling products.

G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á
/50
41
Table 2
Silylformylation of 1-hexyne catalyzed by rhodium on g-Al
2
O
3
catalysts
a
Conversion (%)
b
Catalyst
Rh/g-Al
Rh(TOA)/g-Al
Rh/g-Al (5%) commercial
Time (h)
8
(Z)-1-(Dimethylphenylsilyl)-2-formyl-1-hexene (%)
2
O
3
(1%)
34
100
72
100
100
100
100
2
4
8
0
8
2 3
O (1%)
1
100
O
2 3
Á/
Á/
2
Reaction conditions: Rh catalyst, 0.002 mg atom; toluene, 2 ml; 1-hexyne, 0.23 ml (2 mmol); PhMe SiH, 0.31 ml (2 mmol); carbon monoxide
pressure, 10 atm; 25 8C.
Determined by GLC analysis, based on PhMe
a
2
SiH consumed.
b
Calculated by GLC analysis.
In the presence of additional stabilizers, such as TOA,
the mesitylene-solvated Rh atoms become thermally
stable even at room temperature and they can be more
easily handled in order to deposit Rh nanoparticles on
supports (Scheme 2).
substrate:rhodiumꢂ100. The conversion as function of
time is reported in Fig. 1.
/
It can be seen that, under the adopted reaction
conditions, the commercial rhodium sample is almost
inactive while the MVS-derived rhodium systems are
quite efficient catalysts. However, Rh(TOA)/g-Al O is
The catalytic activity of the so prepared rhodium
systems supported on g-Al O at different Rh loading
has been evaluated and compared with that of a
2
3
more active than the analogous Rh/g-Al O prepared in
2
3
2
3
the absence of the stabilizer.
commercial Rh/g-Al O sample in some significant
2
3
reference reactions such as: (i) the hydrogenation of
methylbenzoate to methylcyclohexanoate, being this
kind of reaction of particular importance in fine
chemical syntheses [1n], (ii) the hydrogenation of
cinnamaldehyde, an a,b-unsaturated aldehyde very use-
ful to evaluate the chemoselectivity of catalysts [11] and
3
.2.2. Hydrogenation of cinnamaldehyde
The hydrogenation of cinnamaldehyde has been
performed at room temperature under atmospheric
pressure of H and Rh(TOA)/g-
using Rh/g-Al
Al catalysts with Rh loadings in the range 0.1Á1%,
lower than those generally required for the hydrogena-
tion of aromatic rings (]5%) [13]. The products,
O
2 3
2
O
3
/
2
(
iii) the silylformylation of 1-hexyne, being the silylfor-
mylation of acetylenic compounds a way to produce
silylalkenals, useful precursors to polyfunctionalized
compounds [12].
/
formed in this reaction, are 3-phenylpropanal (1) and
the cinnamyl alcohol (2) deriving from the hydrogena-
tion of the double bond and the carbonyl group,
respectively, and the 3-phenyl-1-propanol (3) as final
product of the reduction. The results obtained are
summarized in Table 1.
It appears from the reported data that the Rh(TOA)/
g-Al O catalyst is largely more active than the analo-
gous sample prepared in the absence of TOA as
additional stabilizer. Such a difference is much more
evident at very low Rh loading (0.1%) where the
Rh(TOA)/g-Al O catalyst shows a still good activity,
3
3
.2. Catalytic reactions of supported rhodium catalysts
.2.1. Hydrogenation of methyl benzoate
The hydrogenation reaction of methyl benzoate to
2
3
methyl cyclohexanoate with different catalysts (5 wt.%
Rh) has been performed in THF at 80 8C, using a
starting hydrogen pressure of 20 atm and a molar ratio
2
3
Scheme 1.

42
G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á
/50
Scheme 2.
while the Rh/g-Al O system is quite inactive. Under the
3.3. Structural studies of supported rhodium catalysts
2
3
adopted reaction conditions, the commercial Rh/g-
Al O sample is a very poor catalyst. With the exception
In order to rationalize the different behaviour of the
Rh catalysts employed, IR studies of adsorbed CO
species, HRTEM and EXAFS analyses have been
performed on the Rh vapour derived systems Rh/g-
Al O and Rh(TOA)/g-Al O as well as on the com-
2
3
of the 0.1 wt.% Rh/g-Al O system, where high boiling
2
3
unidentified compounds are formed as main products,
all the catalysts show a quite good chemoselectivity in
the hydrogenation of cinnamaldehyde to 3-phenylpro-
panal (2).
2
3
2
3
mercial Rh/g-Al O sample (except EXAFS).
2
3
3
.3.1. IR studies of adsorbed CO
FTIR spectroscopy of adsorbed CO is usually em-
3
.2.3. Silylformylation of 1-hexyne
The silylformylation of acetylenes is an important
ployed to investigate the surface properties of supported
metal particles because of the high sensitivity of the
vibrational characteristics of the surface carbonylic
reaction widely studied using homogeneous Rh catalytic
precursors [1j,14] and no examples are reported in the
literature on silylformylation performed using rhodium-
based heterogeneous catalysts. In general, the reaction
with 1-alkynes gives (Z)-3-silyl-2-alkenals (4) as main
products, although minor amounts of other products,
adducts towards the adsorbing sites properties [1d,15Á
/
18]. In this section, the IR spectra of CO adsorbed on
the Rh/g-Al O catalysts considered, two prepared via
MVS and a commercial one, will be discussed.
Before the admission of CO, the samples were put
2
3
deriving from hydrosilylation, 5, and silylformylationÁ
/
hydrosilylation reactions, 6 and 7, can be detected
(
into contact with 100 Torr H and then outgassed at
2
Scheme 3).
The results obtained in the silylformylation of 1-
room temperature. These conditions were adopted to
obtain information about the surface properties of the
small metal particles in a condition proximal to one in
which they were produced and employed as catalysts
hexyne with CO and dimethylphenylsilane using differ-
ent heterogeneous rhodium on g-Al O catalysts are
reported in Table 2.
The reaction in the presence of the Rh vapour derived
catalysts proceeds with total regio- and stereoselectivity
giving (Z)-1-(dimethylphenylsilyl)-2-formyl-1-hexene as
the sole product, being the Rh(TOA)/g-Al O system
2
3
[
19].
3.3.1.1. IR spectra of CO adsorbed on commercial Rh/g-
Al O . In the case of Rh/g-Al O commercial sample,
the spectrum obtained immediately after admission of
2
3
2
3
2
3
largely more active than the analogous one without
TOA, as also observed in the previously described
hydrogenation reactions. It is noteworthy that the
commercial Rh/g-Al O sample is completely inactive
ꢁ
1
CO shows a main peak at 2065 cm with a shoulder at
ꢁ
1
ca. 2086 cm , a partially resolved component at 2012
ꢁ
1
ꢁ1
cm
and a band centred at 1870 cm
(Fig. 2a). By
2
3
under the adopted reaction conditions.
increasing the contact time with CO up to 2 h (Fig. 2bÁ
/
Scheme 3.

G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á
/50
43
Fig. 2. IR spectra of CO (100 Torr) adsorbed on the commercial Rh/g-
Al (5 wt.% Rh) sample: (a) immediately after CO adsorption; (b)
after 5 min; (c) after 20 min; (d) after 2 h of contact with CO.
2 3
Fig. 3. IR spectra of CO (100 Torr) adsorbed on the Rh/g-Al O
(MVS) (1 wt.% Rh) sample: (a) immediately after CO adsorption; (b)
after 5 min; (c) after 20 min; (d) after 2 h of contact with CO.
2
O
3
d), the intensity of the signals at 2086 and 2012 cmꢁ
progressively increases. On the contrary, the intensity of
1
3), the bands at 2086 and 2012 cmꢁ1 due to Rh(I)(CO)
species were more intense as compared to ones observed
2
ꢁ
1
the bands at 2065 cm slightly decreases and the one at
for the commercial sample (Fig. 2) while the compo-
ꢁ1
ꢁ
1
1
2
870 cm
remain nearly unchanged.
nents due to linear (2043 cm ) and bridged (broad and
1
complex, in the 1900Á
/
1800 cmꢁ range) Rh(0) carbo-
By using a band fitting program, the components at
086 and 2012 cmꢁ1 were extracted from the experi-
mental profile, and their integrated intensities were
calculated. The ratio between the intensity at increasing
time of contact with CO (I ) and the intensity exhibited
nyls were observed as minor components.
The predominance of the components due to
Rh(I)(CO) species indicates that the metal particles in
2
this MVS sample exhibit a high number of surface metal
atoms in low coordination, which are those more easily
extracted by CO to form the volatile Rh(0) carbonyls,
precursors of the Rh(I) dicarbonyls. This fact is in
agreement with the small size of the metal particles of
t
immediately after CO adsorption (I ) was found to be
0
the same for both components, indicating that they are
the same chemical species and, precisely, dicarbonyls on
Rh(I) [20Á23]. The formation of Rh(I)(CO) species by
/
2
interaction of CO with Rh(0) particles has been already
observed in previous works and explained in terms of a
process occurring through the following steps: (i) CO
adsorption on Rh(0) atoms at the surface of rhodium
particles; (ii) extraction of such Rh(0) atoms from the
surface by formation of a volatile rhodium carbonyl (the
exact stoichiometry of which is not known yet); (iii)
migration of the volatile Rh(0) carbonyls to the support,
the hydroxyls of which oxidize the metal atoms to Rh(I),
with the consequent formation of H2.
the Rh/Al O (MVS) sample evidenced by TEM (see
2 3
Fig. 5), as the number of surface atoms in low
coordination on corners and edge is expected to increase
as the size of metal particles decreases [27].
3.3.1.3. IR spectra of CO adsorbed on Rh(TOA)/g-
Al O (MVS). The spectral pattern following CO
2
3
ꢁ
1
Concerning the band at 2065 and 1870 cm , the first
can be assigned to the stretching mode of monocarbo-
nyls linearly adsorbed on Rh(0) particles [24Á
the second is due to carbonyls adsorbed in bridged form
on pairs of rhodium atoms [20Á22].
/26], and
/
3.3.1.2. IR spectra of CO adsorbed on Rh/Al O (MVS).
2 3
Also the spectral pattern obtained by contacting CO
with the Rh/g-Al O (MVS) sample exhibited compo-
2
3
nents due to monocarbonyls (both linear and bridged)
on metal particles and Rh(I) dicarbonyls coming from
the erosion of rhodium atoms. However, significant
differences in relative intensity compared to commercial
sample are evident. In fact, in the spectra recorded
immediately after CO admission at different times (Fig.
Fig. 4. IR spectra of CO (100 Torr) adsorbed on the Rh(TOA)/g-
Al (MVS) (1 wt.% Rh) sample: (a) immediately after CO
2
O
3
adsorption; (b) after 5 min; (c) after 20 min; (d) after 2 h of contact
with CO.

44
G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á50
/
admission on Rh(TOA)/g-Al O (MVS) was found to be
2
3
different from previous samples, in terms of relative
intensity and evolution during the time of the bands due
to the Rh(0) and Rh(I) carbonyls (Fig. 4). The spectrum
recorded immediately after admission of CO onto the
sample (Fig. 4a) showed as main components a band at
ꢁ
1
2
cm
043 cm
and the broad absorption in the 1900Á
/1800
ꢁ
1
range due to linear and bridged monocarbonyls
on Rh(0) at the surface of metal particles, respectively.
ꢁ
1
The signals at 2086 and 2012 cm , due to Rh(I)(CO)2
species, resulting from the surface erosion of the metal
particles, appeared just as two very weak shoulders of
ꢁ
1
the peak at 2043 cm
(Fig. 4a), in sharp contrast with
the situation of Rh/g-Al O (MVS) where they were by
2
3
far the dominant peaks (Fig. 3).
By prolonging the contact time with CO (Fig. 4b and
ꢁ
1
c), even if the shoulders at 2086 and 2012 cm
appeared slightly more pronounced, the mean spectro-
scopic features are those related to carbonyls on Rh(0).
This result appears quite surprising as considering
that analysis TEM showed that in the Rh(TOA)/g-
Al O (MVS) sample, the metal particles are smaller
2
3
than the Rh/g-Al O (MVS) one (see Section 3.3.2), and
2
3
then they should expose at their surface an even higher
number of Rh atoms in low coordination, easily
Fig. 5. Section A: Electron micrographs (original magnification
00,000ꢃ) of commercial Rh/g-Al (A1), Rh/g-Al (MVS) (A2)
and Rh(TOA)/g-Al (MVS) (A3). Section B: Histograms of the
particle size distribution of the samples: commercial Rh/g-Al (B1);
Rh/g-Al (MVS) (B2); Rh(TOA)/g-Al (MVS) (B3).
2
/
O
2 3
2 3
O
abstracted by CO to form Rh(I)(CO) species stabilized
2
2
O
3
on the support. It can be supposed that the inertness
towards this process exhibited by the metal particles of
the Rh(TOA)/g-Al O (MVS) sample resulted from the
2
O
3
2
O
3
2 3
O
2
3
presence of TOA molecules. As the nitrogen is a donor
atom, it could stabilize in some entity the Rh sites
slowing down the process of surface erosion.
Concerning the Rh(TOA)/g-Al
O (MVS) system
2 3
(micrograph in Fig. 5, A3), the metal particle size
distribution of the supported metallic phase, reported
in Fig. 5, B3, appeared further shifted to smaller
diameters, with the resulting mean diameter d ꢂ
/
1.19
/
m
3
(
.3.2. High-resolution transmission electron microscopy
HRTEM) studies
The metal particle size distribution studied by
0.25 nm. Such a distribution is highly asymmetric and
the maximum corresponds to the smallest particle size
detectable by HRTEM. This feature allows us to suggest
the presence of rhodium particles even smaller. In the
end, the presence of a nucleophilic stabilizer such as
TOA during the MVS synthesis, as indicated by the data
obtained, yielded the formation of the smallest metal
particles.
HRTEM of the commercial Rh/g-Al O (5 wt.% Rh)
2
3
system (Fig. 5, A1) shows the presence of metal particles
with a broad size distribution lying in the 1.5Á14 nm
range, with a resulting average diameter of 7.090.5 nm
Fig. 5, B1).
The HRTEM analysis of the Rh/g-Al O (MVS)
/
/
(
2
3
catalyst (micrograph shown in Fig. 5, A2) indicated
the presence of supported rhodium particles smaller
than the commercial sample, having a size distribution
3.3.3. EXAFS studies on the MVS derived catalysts
The EXAFS data have been analysed using the
GNXAS package which allows deconvolutions of the
contributions to the EXAFS signal, due to the local
atomic arrangement around the absorbing atom, in
in a narrow range, 1.0Á/4.0 nm, and a mean diameter
d ꢂ2.190.25 nm (Fig. 5, B2). In this case, moreover,
/
/
m
(
n)
two additional indications has to be considered: the
former is that HRTEM has not an high sensitivity in
revealing particles smaller than 1 nm in size, and the
latter is that the occurrence of aggregation phenomena
of sub-nanometric metal particles under the electron
beam during HRTEM observation cannot be excluded
terms of irreducible n-body contributions, g [30,31].
The EXAFS data analysis by GNXAS has been applied,
as first, to the two reference samples: a-Rh O and Rh
2
3
foil. In Fig. 6, the best-fit results obtained on the total
EXAFS signals and their residuals are reported. The
structural results obtained by the curve-fitting proce-
dures are reported in Table 3. The values found for the
reference samples are in good agreement with the
[28]. So, as observed in previous studies [29], the real
mean size of Rh particles could be even smaller.

G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á
/50
45
2 3
Fig. 6. Best-fit results for Rh foil (left panel) and Rh O (right panel).
expected ones and this allowed us to use the same local
MVS derived catalysts, Rh/g-Al O and Rh(TOA)/g-
2
3
(n)
potentials to generate the g contributions for the
EXAFS signals of the two catalytic samples.
The experimental k-weighted EXAFS signal of rho-
dium oxide was well reproduced using only three single
Al O , are reported in Fig. 7.
2 3
The FT of Rh
first relative to the RhÃ
second includes two contributions (RhÃ
The Rh foil FT spectrum is typical of an FCC structure
and the first four RhÃRh contributions are clearly
visible.
O
shows only two main peaks, the
O first shell distance while the
Rh and RhÃO).
2
3
/
/
/
(
2)
scattering contributions, g , related to the first RhÃ
the second RhÃRh and the third RhÃO coordination
shells. In agreement with the analysis previously re-
/
O,
/
/
/
(
2)
ported [32], the rhodium foil required three g due to
(3)
The Fourier transforms of the two catalysts are quite
different. They present two main peaks that, by
comparison with the model compound spectra, can be
the first three RhÃ
/
Rh coordination shells but also a g
due to the focussing effect which enhances the multiple
scattering contribution to the fourth FCC shell.
The k-weighted Fourier transforms (FTs) of the x(k)
spectra of the two reference samples and of the two
qualitatively assigned to a metalÃ
atom coordination and to a metalÃ
to form small Rh clusters. The analysis and the
/
oxygen or metalÃ
/
light
/
metal coordination,
Table 3
EXAFS structural parameters obtained fitting the reference samples (a-Rh
2
O
3
2 3
and Rh foil) and the Rh/g-Al O sample taking into account an oxide
and a metal phase
Oxide phase
a-Rh
2
O
3
Rh/Al
2
O
3
Metal phase
Rh-foil
2 3
Rh/Al O
N
RhÃO
6.0 (0.3)
2.04 (0.02)
3.0 (0.2)
2.7 (0.3)
2.05 (0.02)
1.3 (0.2)
N
1
12.0 (0.3)
2.68 (0.01)
2.8 (0.2)
4.0 (0.3)
2.69 (0.02)
5.0 (0.2)
1.8 (0.2)
3.80 (0.02)
7.6 (0.3)
5.5 (0.3)
4.69 (0.02)
8.0 (0.3)
˚
˚
1
R (A)
R
RhÃO (A)
ꢁ3
2
2
A )
2
1
ꢁ3
2
A )
˚
˚
s
RhÃO (10
/
/
/
/
(10
N
RhÃRh
3.0 (0.2)
3.01 (0.03)
9.0 (0.2)
0.9 (0.2)
3.01 (0.03)
9.8 (0.2)
N
2
6.0 (0.2)
3.80 (0.01)
2.8 (0.2)
˚
˚
R
RhÃRh (A)
ꢁ3
R
2
(A)
2
2
˚
2
ꢁ3
2
˚
s
RhÃRh(10
A )
(10
A )
2
N
RhÃO
3.0 (0.2)
3.47 (0.03)
20.0 (0.2)
1.2 (0.3)
3.47 (0.03)
14.0 (0.2)
N
3
24.0 (0.4)
4.67 (0.01)
7.5 (0.3)
˚
˚
R
RhÃO (A)
ꢁ3
R
3
(A)
2
2
˚
2
ꢁ3
2
˚
s
RhÃO (10
A )
(10
A )
3
N
4
12.0 (0.3)
5.36 (0.02)
5.7 (0.2)
Á/
˚
R
4
(A)
5.37 (0.03)
9.4 (0.3)
2
ꢁ3
2
˚
A )
(10
u (8)
4
178 (2.0)
183 (3.0)

46
G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á50
/
first three coordination shells coming from Rh O and
2
3
four RhÃ
/
Rh contributions characteristic of metallic
rhodium. As reported in Table 3, the interatomic
distances found are in agreement with the ones of the
reference samples.
For the analysis of the EXAFS data of the Rh(TOA)/
g-Al O sample, different procedures were used. As first
2
3
a situation similar to the one found for the Rh/Al O
2
3
catalyst was taken into account (see Fig. 9) and the
results achieved are reported in Table 4.
In this case, the interatomic distances found for the
second and third shell of the oxide contribution are
different from the a-Rh O ones, while there is a good
2
3
agreement for what concerns the metal contributions.
We conclude that in this case the Rh is not coordinated
as in the a-Rh O oxide phase. It has been well proved
2
3
that in unsupported metal particles stabilized by surfac-
tants, the stabilizer surrounds the metal as a protective
layer [1b,1d,32]; on supported samples although very
few data are at the moment available, it seem reasonable
to suppose that the stabilizer should still cover the metal
particles [1d]. Considering this, a new fitting procedure
was applied assuming a model composed of a metal
phase plus two contributions: RhÃ
/
N and RhÃ
/
C. The
Fig. 7. Fourier transforms, FTs, of the EXAFS data of the reference
samples (Rh foil and Rh ) and of the two MVS derived catalysts
Rh/g-Al and Rh(TOA)/g-Al
best-fit parameters are reported in Table 5 while Fig. 10
shows the comparison between the FTs of the best-fit of
the theoretical model data and of the experimental data.
The good agreement between a theoretical model, in
which the surfactants surround the metal particles, and
the experimental data shows that this model is accep-
table. However, it is not possible to exclude the presence
of other light atoms as nearest neighbours.
2 3
O
O
2 3
2 3
O .
interpretation of the spectra of both catalytic samples is
complicated by the presence of Rh atoms in these two
different phases. In Figs. 8 and 9, the comparison
between best-fit of the EXAFS spectra and of the FTs
of the two catalysts are shown.
In the case of Rh/g-Al O sample, the Rh atoms are
present in two different states: oxide and metal. In fact
best-fit of the data was achieved with the contribution of
Further information can be achieved by EXAFS data
concerning the size of the metallic clusters. Since Rh is
present in two different phases, bound to other Rh
2
3
Fig. 8. The left panels show the experimental kx(k) data (dots) and the best-fit (full line) while the right panels show the comparison of the Fourier
transforms of the kx(k) of the MVS Rh/g-Al catalyst (best-fit is the full line).
2
O
3

G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á
/50
47
Fig. 9. The left panels show the experimental kx(k) data (dots) and the best-fit (full line) while the right panels show the comparison of the Fourier
2 3
transforms of the kx(k) of the MVS Rh(TOA)/g-Al O catalyst (best-fit is the full line).
Table 4
2 3
EXAFS structural parameters obtained fitting Rh (TOA)/g-Al O
sample taking into account an oxide and a metal phase
Oxide phase
Rh (TOA)/
Al _O
Metal phase Rh (TOA)/
Al
2
O
3
2 3
O
N
RhÃO
3.4 (0.3)
2.05 (0.02)
N
1
2.3 (0.2)
2.69 (0.02)
8.8 (0.3)
˚
˚
(A)
R
RhÃO (A)
ꢁ3
R
1
2
2
2
˚
s
RhÃO (10 A ) 1.9 (0.2)
s
1
10ꢁ A )
3
2
˚
(
N
RhÃRh
0.4 (0.1)
3.07 (0.04)
N
2
1.2 (0.2)
3.80 (0.03)
15.3 (0.3)
˚
˚
(A)
R
RhÃRh (A)
ꢁ3
R
2
2
2
2
˚
s
RhÃRh(10 A ) 2.0 (0.4)
s
2
10ꢁ A )
3
2
˚
(
N
RhÃO
2.9 (0.2)
3.22 (0.04)
N
3
3.0 (0.3)
4.70 (0.04)
13.0 (0.4)
˚
˚
(A)
R
RhÃO (A)
ꢁ3
R
3
2
2
2
˚
s
RhÃO (10 A ) 2.5 (0.3)
s
2
10ꢁ A )
3
2
˚
(
Fig. 10. Comparison of the Fourier transforms of the kx(k) of the
catalyst (best-fit is the full line) and of the
theoretical model including N.
2 3
MVS Rh(TOA)/g-Al O
Table 5
EXAFS structural parameters obtained fitting Rh(TOA)/g-Al
a theoretical model that includes only two light elements contributions
2 3
O with
atoms and to light backscattering atoms, it is not
possible to deduce the Rh cluster dimensions in a simple
way, from the coordination numbers (N), as usually
done. In this case, in fact, as shown elsewhere [33], the N
values found are not the real coordination numbers
(
RhÃ
/
N and RhÃ/C) and a metal phase
Rh (TOA)/
Metal phase Rh (TOA)/
Al
Al
2
3
O _N
2 3
O
N
RhÃN
3.7 (0.3)
2.07 (0.02)
3.2 (0.2)
N
1
2.6 (0.3)
2.70 (0.02)
10.5 (0.3)
(
N ) but are weighted by the atomic percentage (a) of
th
˚
˚
(A)
R
RhÃN (A)
R
1
2
2
Rh atoms present in the metallic phases, i.e. NꢂaN .
/
th
s
RhÃN
10 A )
s
1
ꢁ
3
2
ꢁ3
2
Because the atomic percentage, a, is not known, we
calculated the expected theoretical mean coordination
numbers for the first three coordination shells assuming
clusters of a spherical shape with an FCC structure and
˚
˚
(10 A )
(
N
2
1.1 (0.2)
3.80 (0.03)
18.4 (0.3)
˚
(A)
R
2
2
s
2
3
2
˚
(
10ꢁ A )
radius R . Dividing the experimental N values by the
c
N
RhÃC
2.8 (0.2)
3.24 (0.04)
3.1 (0.3)
N
3
2.9 (0.3)
4.70 (0.04)
16.5 (0.4)
˚
˚
(A)
calculated ones, we obtain three estimations of a as a
function of the cluster radius, R , one from each
c
R
RhÃC (A)
R
3
2
2
s
RhÃC
10 A )
s
2
ꢁ
3
2
ꢁ3
2
(10 A )
˚
˚
(
coordination shell.

48
G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á50
/
Fig. 11. Ratio, a, of the experimental and theoretical N values for the Rh/g-Al
the cluster radius, R
O
2 3
2 3
and the MVS Rh(TOA)/g-Al O samples drawn as a function of
c
.
Apart from statistical errors, the three curves ob-
tained should intersect at an R value that represents the
Rh/g-Al O samples. The Rh/g-Al O commercial cata-
2
3
2
3
c
lyst shows, under the adopted reaction conditions, a
actual cluster dimension. From a more realistic point of
view, taking into account the errors, for each shell two
curves are obtained that do not define a point but an
poor catalytic activity, while the Rh(TOA)/g-Al O
2
3
sample, prepared from mesitylene-solvated Rh atoms,
stabilized by TOA, is a very efficient catalyst, resulting
also more active than the analogous Rh/g-Al O sample
overlap region whose R range gives an estimation of the
c
2
3
possible metallic cluster dimension (Fig. 11). For Rh/
˚
prepared in the absence of TOA.
Al O sample, the overlap region is in the range 8.0 AB
/
The structural characterization of the above catalysts
satisfactory accounts of this different behaviour. The IR
spectrum detected after treatment with CO of the Rh/g-
Al O commercial sample shows peaks consistent with
2
3
˚
R B11 A while for the Rh(TOA)/g-Al O sample, the
/
c
2
3
˚
˚
overlap region is in the range 3.5 ABR B5 A. The
mean diameters so determined are: D19 A for Rh/g-
Al O and D8 A for Rh(TOA)/g-Al O .
/
/
c
˚
2
3
˚
2
3
2
3
the presence of Rh(I) and Rh(0) carbonyl species. In the
The mean diameters found using EXAFS are in both
cases smaller than those found with HRTEM. As
already underlined [29,34] in the presence of a cluster
size distribution, EXAFS measures an average dimen-
sion weighted by the fraction of atoms corresponding to
all the different sizes. For this reason, we can conclude
that in both samples there is a large amount of Rh
MVS Rh/g-Al O derived catalyst, the components due
2
3
to the Rh(I) carbonyls are predominant; this result
agrees with the presence of small rhodium metal
particles from which rhodium atoms can be easily
extracted to form the volatile Rh(0) carbonyl precursors
of Rh(I) species (vide infra). The HRTEM data are
consistent with this finding showing the presence of
metal particles averaging 7.0 nm in the commercial
sample and 2.1 nm in the Rh/g-Al O catalyst.
˚
atoms in particles with dimensions smaller than 10 A,
which are clearly visible by EXAFS but are not seen by
HRTEM and that the average dimensions of the clusters
is lower in the presence of TOA than in its absence. It
seems that the main effect of TOA is to prevent the
aggregation of atoms/clusters to form larger particles. It
can be due to the presence of a protective layer around
2
3
The IR spectrum following CO admission on
Rh(TOA)/g-Al O catalyst is really different from those
2
3
of the previous sample. It shows as main component
only bands due to Rh(0) species. This is quite surprising
considering that HRTEM analysis indicates the pre-
sence of very small Rh clusters, averaging 1.1 nm in
diameter which should have more rhodium atoms easily
attacked by CO to form Rh(I) carbonyl species. It is
then reasonable to suppose that TOA still surrounds the
the metal core, which avoids direct metalÃ
/metal con-
tacts.
4
. Conclusions
The results obtained in some important reference
metal clusters deposited on g-Al
erosion.
The EXAFS study strongly supports this hypothesis.
The Fourier transforms of the EXAFS data of the two
O avoiding surface
2 3
reactions, as the hydrogenation of methyl benzoate and
cinnamaldehyde and the silylformylation of 1-hexyne,
have evidenced a quite different catalytic activity of the
MVS derived catalysts, Rh/g-Al
Al O , are quite different. While in the Rh/g-Al O
O
and Rh(TOA)/g-
2
3
2
3
2
3

G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á
/50
49
Schmid, Appl. Organomet. Chem. 15 (2001) 178;
k) H. B o¨ nnemann, R. Richards, Eur. J. Inorg. Chem. (2001)
455.;
l) L.A. Aronica, S. Terreni, A.M. Caporusso, P. Salvadori, Eur.
sample, the Rh atoms are present in two different states,
(
that is as oxide and as metal, in the Rh(TOA)/g-Al O
2
3
2
sample the Rh atoms are not present as an oxide phase
and the experiment data are in good agreement with a
theoretical model composed of a metal phase plus a
contribution due to the presence of the amine and
(
J. Org. Chem. (2001) 4321.;
(m) G. Vitulli, M. Bernini, S. Bertozzi, E Pitzalis, P. Salvadori, S.
Coluccia, G. Martra, Chem. Mater. 14 (2002) 1183;
(
n) J.A. Widegren, R.G. Finke, J. Mol. Catal. A 191 (2003) 187.
attributed to RhÃ
/
N or RhÃ
/
C interactions.
[
2] (a) See as examples: S. Abbet, A. Sanchez, U. Heiz, W.-D.
Schneider, A.M. Ferrari, G. Pacchioni, N.Rosch, J. Am. Chem.
Soc., 122, 2000, 3453, and references therein.;
The enhanced activity of the Rh(TOA)/g-Al O3
2
catalyst can be then reasonably related with these
factors. In fact, the presence of homogeneously distrib-
uted metal particles of largely small size makes available
more rhodium atoms for the catalysis. In addition, the
stabilization action of TOA on the particles prevents
poisoning and an extensive oxidation by the solid oxide
support. They can also satisfactorily explain why
catalysts at very low Rh loading (0.1%) are still active
in the hydrogenation of cinnamaldehyde only when
prepared in the presence of TOA (Table 1).
(b) R. Rao, G.U. Kulkarni, P.J. Thomas, P.P. Edwards, Chem.
Eur. J. 8 (2002) 29.
3] (a) G. Schmid, Chem. Rev. 92 (1992) 1709;
[
[
(
(
b) H. Hirai, H. Chawanya, N. Toshima, Reactive Polym. 3
1985) 127.
4] V.M. Frolov, Platinum Met. Rev. 40 (1996) 8.
[5] G. Vitulli, P. Pertici, S. Bertozzi, A.M. Caporusso, R. Lazzaroni,
P. Salvadori, in: W.A. Herrmann (Ed.), Synthetic Methods of
Organometallic and Inorganic Chemistry: Catalysis, Brauer/
Herrmann, Munich, 2002, p. 10.
[
6] G. Vitulli, E. Pitzalis, R. Lazzaroni, P. Pertici, P. Salvadori, O.
These catalysts can be used several times without
valuable lost of activity. In preliminary experiments, it
has been observed that hexadecylamine and trioctyla-
minomethylamonium bromide behave like TOA, while,
using polyvinylpyrrolidone, the resulting catalyst exhibit
a quite low activity, probably because of a more difficult
access of substrates to the metal surface.
Salvetti, S. Coluccia, G. Martra, Mater. Sci. Forum 195 (1995) 93.
[7] S. Pascarelli, F. D’Acapito, G. Antonioli, A. Balerna, F.
Boscherini, R. Cimino, G. Dalba, P. Fornasini, G. Licheri, C.
Meneghini, F. Rocca, S. Mobilio, ESRF Newslett. 23 (1995) 17.
[
8] S. Pascarelli, F. Boscherini, F. D’Acapito, J. Hrdy, C. Meneghini,
S. Mobilio, J. Synch. Rad. 3 (1996) 147.
[
9] G. Vitulli, G. Uccello-Barretta, P. Pannocchia, A. Raffaelli, J.
Organomet. Chem. 302 (1986) C21.
[10] G. Vitulli, E. Pitzalis, P. Pertici, P. Salvadori, S. Coluccia, G.
Martra, L. Lampugnani, M. Mascherpa, Mater. Sci. Eng. C 15
(
2001) 207.
Acknowledgements
[
[
11] B. Coq, F. Figueras, Coord. Chem. Rev. 178Á180 (1998) 1753.
12] I. Ojima, M. Tzamarioudaki, C.Y. Tsai, J. Am. Chem. Soc. 116
(1994) 3643.
/
We gratefully acknowledge financial supports by
CNR, Italy (PF-MSTA II) and MIUR, Italy (low 95/
[13] See as example: P.N. Rylander, Hydrogenation Methods, 1985,
17, Academic Press, New York.
1
14] I. Ojima, Chem. Rev. 88 (1988) 1011.
9
5).
[
[
15] See as example: L. Marchese, G. Martra, S. Coluccia, C.R.A.
Catlow, A. Cheetham, New Trends in Materials Chemistry, 1997,
p. 79, Kluwer Academic Publishers, London.
References
[16] C. Louis, L. Marchese, S. Coluccia, A. Zecchina, J. Chem. Soc.
[
1] (a) K.J. Klabunde, Free Atoms, Clusters and Nanoscale Particles,
Accademic Press, New York, 1994;
Faraday Trans. 185 (1989) 1655.
[17] A. Parmigliana, F. Arena, F. Fruster, S. Coluccia, L. Marchese,
G. Martra, A.L. Chuvilin, J. Catal. 141 (1993) 34.
[18] G. Martra, H.M. Swaan, C. Mirodatos, M. Kermarec, C. Louis,
Stud. Surf. Sci. Catal. 111 (1997) 617.
(
b) N. Nasar, F. Fache, M. Lemaire, J.-C. Beziat, M. Besson, P.
Gallezot, J. Mol. Catal. 87 (1994) 107;
c) G. Vitulli, A. Verrazzani, A.M. Caporusso, E. Pitzalis, P.
Pertici, P. Salvadori, in: S. Daolio, E. Tondello, P.F. Vigato
Eds.), Syntheses and Methodologies in Inorganic Chemistry:
New Compounds and Materials, vol. 7, La Photograph, Padova,
997, p. 52;
d) H. B o¨ nnemann, G. Braun, W. Brijoux, R. Brinkmann, A.
Schulze Tilling, K. Seevogel, K. Siepen, J. Organomet. Chem. 520
(
[19] D. Martim, D. Duprez, Appl. Catal. A 131 (1995) 297.
[20] H.F.J. van’t Blik, J.B.A.D. van Zon, T. Huizinga, J.C. Vis, D.C.
Koningsberger, R. Prins, J. Am. Chem. Soc. 107 (1985) 3139.
[21] D.K. Paul, C.D. Marten, Langmuir 15 (1999) 4508.
[22] N. Sheppard, T.T. Nguyen, in: R.J.H. Clark, R.E. Hester (Eds.),
Advanced Infrared Spectroscopy of Adsorbed Species, vol. 5,
Heyden, London, 1978, p. 67 (Chapter 2).
(
1
(
(
(
1996) 143;
e) Y. Wang, H. Liu, N. Toshima, J. Phys. Chem. 100 (1996)
[23] K.A. Almusaiteer, S.S.C. Chuang, J. Phys. Chem. B 104 (2000)
2265.
1
9533;
f) G. Vitulli, E. Pitzalis, A. Verrazzani, P. Pertici, P. Salvadori,
G. Martra, Mat. Sci. Forum 929 (1997) 235;
(
[24] P. Basu, D. Panayotov, J.T. Yates, Jr., J. Am. Chem. Soc. 110
(1988) 2074.
(
(
(
g) K.S. Weddle, J. Aiken, III, R.G. Finke, J. Am. Chem. Soc. 120
1998) 5653;
h) N. Toshima, T. Yonazawa, New J. Chem. 22 (1998)
[25] E.C. Decanio, D.A. Storm, J. Catal. 132 (1991) 375.
[26] T.S. Zubkov, E.A. Wovchko, J.T. Yates, Jr., J. Phys. Chem. B
103 (1999) 5300.
1
179;
i) J. Schulz, A. Roucoux, H. Patin, JCS Chem. Commun. (1999)
35.;
j) N. Toshima, Y. Shiraishi, T. Teranishi, M. Miyake, T.
Tominaga, H. Watanabe, W. Brijoux, H. B o¨ nnemann, G.
[27] M. Che, C.O. Bennett, Adv. Catal. 36 (1989) 55.
[28] G. Martra, S. Coluccia, O. Monticelli, G. Vitulli, Catal. Lett. 29
(1994) 105.
(
5
(
[29] A. Balerna, S. Coluccia, G. Deganello, A. Longo, A. Martorana,
G. Martra, C. Meneghini, P. Pertici, G. Pipitone, E. Pitzalis, A.M.

50
G. Vitulli et al. / Journal of Organometallic Chemistry 681 (2003) 37Á50
/
Venezia, A. Verrazzani, G. Vitulli, Eur. Phys. J. D 7 (1999)
77.
30] A. Filipponi, A. Di Cicco, S.R. Natoli, Phys. Rev. B 5 (1995)
5122.
(b) T. Yonezawa, T. Tominaga, R. Dominique, J. Chem. Soc.
Dalton Trans. (1996) 783.
5
[
[33] A. Balerna, A. Longo, A. Martorana, C. Meneghini, S. Mobilio,
G. Pipitone, Eur. Phys. J. D 7 (1999) 89.
1
[
31] A. Filipponi, A. Di Cicco, Phys. Rev. B 5 (1995) 15135.
32] (a) M.T. Reetz, W. Helbig, S.A. Quaiser, U. Stimming, N. Breuer,
R. Vogel, Science 267 (1995) 367;
[34] S. Polizzi, P. Riello, A. Balerna, A. Benedetti, Phys. Chem. Chem.
Phys. 3 (2001) 4614.
[